Skip to Main Content
Behavior Analysis in visual surveillance has become a very active issue for the computer vision research community, particularly when crowded scenes are concerned. In this perspective, motion analysis and tracking is challenging due to significant visual ambiguities which incite to look into more alternative solutions. In this paper we introduce a new framework for recognizing various motion patterns, extracting abnormal behaviors and tracking them over crowded traffic scenes. The proposed approach highlights three traffic density levels and performs in two modes: an “off-line” mode for motion patterns learning and modeling, and an “on-line” mode for distinguishing irregular motions and tracking them separately.